Pith. sign in

REVIEW 4 major objections 6 minor 32 references

TO-GATE: Clarifying Questions and Summarizing Responses with Trajectory Optimization for Eliciting Human Preference

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read TO-GATE trains LLMs to ask better clarifying questions by optimizing whole dialogue trajectories with a DPO-style contrastive loss, beating prior methods by 9.32%.

desk verdict A plausible incremental training recipe for preference elicitation, but the evidence as reported is too shaky to accept: single-run small evaluation, a circular clarification metric, and an internal lambda inconsistency. read the letter →

arxiv 2506.02827 v1 pith:4WOGZEKW submitted 2025-06-03 cs.CL

Yulin Dou , Jiangming Liu This is my paper
classification cs.CL
keywords preferenceelicitationclarifyingquestionstrajectoryoptimizationDirectdialogueagentspersonalizedresponsegenerationmulti-turnLLMalignment
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that learning to ask clarifying questions and to summarize a final personalized answer works best when the model optimizes whole dialogue trajectories, not when it merely copies high-scoring conversations. Its TO-GATE framework combines a clarification resolver, which uses contrastive preference learning on positive and negative question paths, with a summarizer that separately weights the loss on final responses. On standard preference-elicitation tasks, the trained questioner beats STaR-GATE, the supervised positive-trajectory baseline, with an average win rate of 83.15 versus 73.83. The authors argue that explicitly penalizing degenerate dialogues and separating the learning signal for questions from the signal for final answers is what drives the improvement.

What carries the argument

The load-bearing mechanism is trajectory optimization with two coupled modules. The clarification resolver repeatedly explores dialogues with a roleplayer, selects the best and worst conversations by the frozen base model's log-likelihood of the gold response, and applies the DPO contrastive loss to prefer the good trajectory over the bad one, with a dynamic dataset that grows through an exploration-collection-training loop. The summarizer adds a separate DPO term for the final response, combined with the clarification loss through a weight parameter λ. A deterministic dual-pass evaluation, which averages win rates over both response orderings, is introduced to remove position bias from the GPT-4 judgments.

What would settle it

If a different frozen base model or a different gold-response distribution changed which dialogues are selected as best and worst, and the reported win-rate advantage shrank or reversed, the filtering metric would be shown to be the active ingredient rather than the training method. A direct test would be to train TO-GATE with trajectories selected by human judges or by an oracle reward, and compare win rates against the likelihood-selected version.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that treating the entire clarifying dialogue as a trajectory and applying Direct Preference Optimization-style contrastive learning to distinguish successful from unsuccessful trajectories yields better questions and final responses than supervised fine-tuning on the best trajectories alone. The improvement comes from two coupled ingredients: penalizing bad conversation paths, and using a separate loss term for the final response with an adjustable weight. The paper reports a 9.32% improvement over the strongest baseline and state-of-the-art results on GATE-style preference elicitation tasks, with the clarification resolver contributing the larger share of the gain.

Load-bearing premise

The paper assumes that the frozen base model's likelihood of producing the gold response after a dialogue is a faithful measure of how well that dialogue elicits human preferences, and the same measure both filters training data and scores the results.

Editorial extensions

If this is right

  • Question-asking policies improve from contrastive trajectory learning: TO-GATE's 83.15 average win rate beats STaR-GATE's 73.83.
  • Penalizing bad trajectories matters: removing the clarification resolver lowers win rate by 5.65 percentage points, more than removing the summarizer, which costs 2.12 points.
  • Separate losses for questions and final responses improve alignment: a λ near 2, which balances the two terms, outperforms settings that over-weight responses.
  • Averaging over both response orderings in evaluation removes position bias and gives reproducible comparison scores.
  • The trend of the curves suggests that additional training epochs could further improve the model, since clarification performance keeps rising through M3.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: the evaluation metric and the training filter are the same quantity, log p_QBASE(gold response | dialogue), so the reported clarification gains may partly reflect the model learning to satisfy that particular likelihood signal rather than elicitation quality in general; a human study or a held-out judge would test this.
  • Inference: the trajectory-optimization loop is not specific to preference elicitation and could transfer to other multi-turn interaction tasks where a gold outcome or user-satisfaction signal is available.
  • Inference: because the authors note that M3 shows a divergence between clarification and response trends, an interesting extension would be an early-stopping or adaptive weighting schedule keyed to response quality rather than clarification quality alone.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes TO-GATE, a framework for training an LLM questioner to elicit human preferences through multi-turn clarifying questions and then summarize them into a final personalized response. The method frames this as trajectory optimization: for each task, it samples candidate dialogues, selects the best and worst trajectories according to the base model's likelihood of the gold response, and then trains with a DPO-style contrastive loss that is split into a clarification loss and a response loss with weight lambda. Experiments on a synthetic GATE-style benchmark compare TO-GATE with STaR-GATE and DPO, reporting win rates against the untrained base model and a clarification-score curve. The paper claims state-of-the-art performance, with a 9.32% improvement over STaR-GATE.

Significance. If the empirical claims held, TO-GATE would be a useful extension of STaR-GATE: replacing supervised training on positive trajectories with contrastive training over dynamically filtered trajectories is a plausible way to penalize uninformative clarification dialogues. The deterministic dual-pass evaluation for win rates is a reasonable safeguard against position bias and is a methodological strength. The DPO derivation in Section 4.1 is standard and correctly adapted to the trajectory setting. However, the paper's central empirical claim is not currently substantiated: the headline result rests on a single run, the clarification metric is the same quantity used for training-data selection, the closest DPO-based baselines are described but not compared, and the lambda used in the main experiments is inconsistent with the tuning study. These issues are load-bearing because the paper's contribution is primarily empirical.

major comments (4)
  1. [§5.4, Table 1, Appendix B] The headline response result is a single run on a test set of 10 personas × 50 tasks, with no standard deviation, confidence interval, or significance test. The 9.32-percentage-point gap between TO-GATE (83.15) and STaR-GATE (73.83) is therefore not established as real; it may be within run-to-run and GPT-4 judge variation. Moreover, each row of Table 1 is a win rate against M0, so the difference between rows is an indirect comparison, not a pairwise test. Please report multiple random seeds (at least 3) and paired significance tests over the 50 tasks.
  2. [§4.4 (Eq. 11) and Algorithm 1 (line 4)] The clarification metric in Eq. (11) is the log-probability the base model assigns to the gold response given the dialogue. Algorithm 1 line 4 selects the 'best' and 'worst' trajectories using the same quantity (with π0 in place of QBASE). Training therefore directly optimizes the quantity used to report clarification improvements, so the monotone improvement in Figure 4 is partly by construction and does not provide independent evidence of better preference elicitation. Please evaluate clarification quality with an independent measure, or at least discuss this coupling and show that response win rates improve without relying on the trained metric.
  3. [§5.3 and §5.5 (Table 2)] The main experiments set λ=0.33, but the only λ sweep reports λ∈{1,2,3,6} and finds λ=2 best at 78.66%, well below the 83.15% reported in Table 1. The paper does not explain why λ=0.33 was selected or how it relates to the sweep, and it does not report the λ=0.33 point. This inconsistency makes the reported configuration hard to reproduce and weakens the claim that the chosen hyperparameters are well motivated.
  4. [§2 Related Work and §5.2 Baselines] The related work identifies ETO (Song et al., 2024) as an extension of DPO to multi-turn dialogues and cites DMPO and SDPO, but none of these methods appears in the experiments. Because ETO is the closest comparison for trajectory-level DPO, omitting it leaves the 'state-of-the-art' claim incomplete. Please include ETO (and ideally DMPO) under the same evaluation protocol.
minor comments (6)
  1. [Figure 4] The x-axis of Figure 4 shows M0–M4 while Section 5.2 says all models are trained for three epochs; Figure 5 uses M0–M3. Please reconcile the number of epochs displayed and discussed.
  2. [Throughout] There are numerous typos: 'contrastive learing' in the Figure 1 caption, 'looses' after Eq. (10), 'Reloplayer' in Table 3, and 'tend of reults curves' in Section 5.4.
  3. [Abstract] The abstract contains a stray '}' after the STaR-GATE citation; please fix the LaTeX error.
  4. [Eq. (1)] Equation (1) uses pQBASE before QBASE is defined; please define QBASE before first use and clarify its relationship to π0 in Algorithm 1.
  5. [§4.4] The text says the evaluation 'averages scores across all possible response orderings', but only two orderings are possible for a pair; consider rephrasing to 'both orderings'.
  6. [Table 1] The column headers A-B and B-A are not defined in the caption; please specify which model is A and which is B in the dual-pass evaluation.

Circularity Check

1 steps flagged · score 6.0 of 10

Partial circularity: the clarification metric (Eq. 11) is the same base-model gold-likelihood used to select positive/negative trajectories for DPO training (Algorithm 1 line 4), so the reported monotone clarification improvement is partly forced by construction; the headline response win-rate claim is not itself circular.

  1. fitted input called prediction [Section 4.4, Eq. (11); Algorithm 1, line 4; Section 3, Eq. (1)]
    "We use the log-probability of gold responses og ij conditioning on the simulated dialogue history sij given by Q: log pQBASE(og ij | ti, sij), (11) where QBASE is a base language model without training. ... Select best and worst conversations by base policy likelihood: sw ij = arg maxsc ij log pπ0(og ij|ti, sc ij), s l ij = arg minsc ij log pπ0(og ij|ti, sc ij)"

    The 'clarification' result is measured by the frozen base model's log-likelihood of the gold response given the dialogue, p_QBASE(gold | s). The training pipeline selects positive and negative dialogues by exactly the same functional form, p_π0(gold | s), with π0 the untrained base policy. DPO then trains the questioner to imitate high-scoring dialogues and avoid low-scoring ones. The evaluation therefore reports improvement on the very quantity used to filter the training data, so the monotone clarification gain in Figure 4 is at least partly an artifact of optimizing the metric through data selection rather than an independent measure of preference-elicitation quality.

full rationale

One concrete circular coupling exists: the clarification-quality metric of Eq. (11) is the same base-model gold-likelihood used in Algorithm 1 line 4 to construct the positive/negative trajectory pairs for DPO training. Because the training signal is derived from the evaluator's own score, the reported superiority of TO-GATE on the clarification axis (Figure 4) is partly forced by construction. This warrants a 6 rather than a lower score. However, the paper's headline '9.32% improvement' and the response win rates in Table 1 are based on a separate GPT-4 judgment protocol, not on the base-model likelihood, so the central response-level claim is not circular in the same way; its weaknesses (single run, no significance test, no ETO comparison, λ inconsistency) are statistical and reproducibility concerns, not circularity. I found no load-bearing self-citations: the authors do not cite their own prior work as justification, and the DPO derivation follows standard Rafailov et al. math. The oracle/roleplayer/judge all being LLMs is a validity threat but not a definitional reduction.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on three domain assumptions that are not independently validated: the likelihood-based clarification metric, the fidelity of the GPT-4 and roleplayer simulation, and the bias removal by ordering averaging. The free parameters lambda and beta are hand-set or tuned on the evaluation set. No new physical or conceptual entities are introduced.

free parameters (3)
  • lambda (lambda) = 0.33 in main runs; 2 selected as best in Table 2
    Weights the final-response DPO loss relative to the clarification loss (Eq. 10). Reported inconsistently: Section 5.3 says lambda=0.33, while Table 2 tunes lambda and finds lambda=2 best. This is a hand-chosen hyperparameter affecting the final win rate.
  • beta (beta) = 0.1
    DPO temperature controlling preference strength and KL regularization (Sections 4.1 and 5.3). Fixed by hand without sensitivity analysis.
  • number of sampled trajectories per task = 10
    Algorithm 1 line 3 samples 10 conversations per (task, persona) and selects the best and worst by base-policy likelihood. The choice of 10 is not justified and affects which trajectories become positives and negatives.
assumptions (4)
  • domain assumption Log-likelihood of the oracle gold response given a dialogue is a valid measure of clarification quality.
    Used in Eq. (11) as the clarification metric and in Algorithm 1 line 4 to filter training trajectories. No independent evidence connects this likelihood to human preference elicitation quality.
  • standard math Bradley-Terry model and DPO reward reparameterization apply to multi-turn clarification trajectories.
    Equations (4)-(8) follow the standard DPO derivation from Rafailov et al. (2023); this is accepted background.
  • domain assumption GPT-4 as Oracle and judge, and the Mixtral roleplayer, faithfully represent real user personas and preferences.
    The entire training and evaluation loop is simulated. The Limitations section warns that real-world preference elicitation may differ from these simulations.
  • domain assumption Averaging the two response orderings removes GPT-4 position bias.
    The dual-pass evaluation averages A-B and B-A scores, but no analysis shows the remaining bias is symmetric or negligible.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TO-GATE: Clarifying Questions and Summarizing Responses with Trajectory Optimization for Eliciting Human Preference." pith.science (2026). https://pith.science/paper/4WOGZEKW

@misc{pith2026250602827,
  author       = {Pith},
  title        = {Pith review of: TO-GATE: Clarifying Questions and Summarizing Responses with Trajectory Optimization for Eliciting Human Preference},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4WOGZEKW}},
  note         = {Machine review of arXiv:2506.02827}
}
read the original abstract

Large language models (LLMs) can effectively elicit human preferences through multi-turn dialogue. Complex tasks can be accomplished through iterative clarifying questions and final responses generated by an LLM acting as a questioner (STaR-GATE; Andukuri et al., 2024}). However, existing approaches based on self-taught reasoning struggle to identify optimal dialogue trajectories and avoid irrelevant questions to the tasks. To address this limitation, we propose TO-GATE, a novel framework that enhances question generation through trajectory optimization, which consists of two key components: a clarification resolver that generates optimal questioning trajectories, and a summarizer that ensures task-aligned final responses. The trajectory optimization enables the model to produce effective elicitation questions and summary responses tailored to specific tasks. Experimental results demonstrate that TO-GATE significantly outperforms baseline methods, achieving a 9.32% improvement on standard preference elicitation tasks.

Figures

Figures reproduced from arXiv: 2506.02827 by the authors.

Figure 1
Figure 1. (a) STaR-GATE adopts the supervised train [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. In the training phase, TO-GATE optimizes the clarification resolver by using contrastive learning to [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The dynamic dataset Dp is generated and updated through a continuous cycle of “exploration￾collection-training,” enabling dynamic expansion of the training data and continuous improvement of model performance. is formulated as: max πθ E(ti,uj )∼D,qij∼πθ(ti|qij ) [r(ti , qij )]− βDKL [πθ(qij | ti) ∥ πref(qij | ti)] , (3) where, πθ and πref denotes the policy model (e.g., Questioner), and qij represents a clarificatio… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Results of clarification and responses across [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Results of responses given by models with [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: The prompt designed to guide the Questioner’s [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: The prompt designed to guide the Roleplayer [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: The prompt for generating personalized and [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

32 extracted references · 15 canonical work pages

  1. [1]

    Mohammad Aliannejadi, Julia Kiseleva, Aleksandr Chuklin, Jeffrey Dalton, and Mikhail Burtsev. 2021. Building and evaluating open-domain dialogue corpora with clarifying questions. arXiv preprint arXiv:2109.05794

    work page Pith review arXiv 2021

  2. [2]

    Chinmaya Andukuri, Jan-Philipp Fr \"a nken, Tobias Gerstenberg, and Noah Goodman. 2024. Star-gate: Teaching language models to ask clarifying questions. In First Conference on Language Modeling (COLM)

    work page 2024

  3. [3]

    Yuntao Bai, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, and 1 others. 2022. Constitutional ai: Harmlessness from ai feedback. arXiv preprint arXiv:2212.08073

    arXiv 2022

  4. [4]

    Tin Barisin, Katja Schladitz, and Claudia Redenbach. 2024. Riesz networks: Scale-invariant neural networks in a single forward pass. Journal of Mathematical Imaging and Vision, 66(3):246--270

    work page 2024

  5. [5]

    Ralph Allan Bradley and Milton E Terry. 1952. Rank analysis of incomplete block designs: I. the method of paired comparisons. Biometrika, 39(3/4):324--345

    1952

  6. [6]

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, and 1 others. 2020. Language models are few-shot learners. Advances in neural information processing systems, 33:1877--1901

    2020

  7. [7]

    Chelsea Finn, Kelvin Xu, and Sergey Levine. 2018. Probabilistic model-agnostic meta-learning. Advances in neural information processing systems (NeurIPS), 31

    work page 2018

  8. [8]

    Jan-Philipp Fr \"a nken, Sam Kwok, Peixuan Ye, Kanishk Gandhi, Dilip Arumugam, Jared Moore, Alex Tamkin, Tobias Gerstenberg, and Noah D Goodman. 2023. Social contract ai: Aligning ai assistants with implicit group norms. arXiv preprint arXiv:2310.17769

    arXiv 2023

Show all 32 references
  1. [9]

    Louie Giray. 2023. Prompt engineering with chatgpt: a guide for academic writers. Annals of biomedical engineering, 51(12):2629--2633

    2023

  2. [10]

    Kunal Handa, Yarin Gal, Ellie Pavlick, Noah Goodman, Jacob Andreas, Alex Tamkin, and Belinda Z Li. 2024. Bayesian preference elicitation with language models. arXiv preprint arXiv:2403.05534

    2024 arXiv

  3. [11]

    Joey Hong, Sergey Levine, and Anca Dragan. 2023. Zero-shot goal-directed dialogue via rl on imagined conversations. arXiv preprint arXiv:2311.05584

    2023 arXiv

  4. [12]

    Arian Hosseini, Xingdi Yuan, Nikolay Malkin, Aaron Courville, Alessandro Sordoni, and Rishabh Agarwal. 2024. V-star: Training verifiers for self-taught reasoners. arXiv preprint arXiv:2402.06457

    2024 arXiv

  5. [13]

    Aobo Kong, Wentao Ma, Shiwan Zhao, Yongbin Li, Yuchuan Wu, Ke Wang, Xiaoqian Liu, Qicheng Li, Yong Qin, and Fei Huang. 2025. Sdpo: Segment-level direct preference optimization for social agents. arXiv preprint arXiv:2501.01821

    2025 arXiv

  6. [14]

    Li, Alex Tamkin, Noah Goodman, and Jacob Andreas

    Belinda Z. Li, Alex Tamkin, Noah Goodman, and Jacob Andreas. 2025. Eliciting human preferences with language models. In The Thirteenth International Conference on Learning Representations (ICLR)

    2025

  7. [15]

    Jessy Lin, Nicholas Tomlin, Jacob Andreas, and Jason Eisner. 2024. Decision-oriented dialogue for human-ai collaboration. Transactions of the Association for Computational Linguistics, 12:892--911

    2024

  8. [16]

    Qidong Liu, Xian Wu, Xiangyu Zhao, Yuanshao Zhu, Derong Xu, Feng Tian, and Yefeng Zheng. 2024. When moe meets llms: Parameter efficient fine-tuning for multi-task medical applications. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in...

    2024

  9. [17]

    Andrea Madotto, Zhaojiang Lin, Genta Indra Winata, and Pascale Fung. 2021. Few-shot bot: Prompt-based learning for dialogue systems. arXiv preprint arXiv:2110.08118

    2021 arXiv

  10. [18]

    Ben Mann, N Ryder, M Subbiah, J Kaplan, P Dhariwal, A Neelakantan, P Shyam, G Sastry, A Askell, S Agarwal, and 1 others. 2020. Language models are few-shot learners. arXiv preprint arXiv:2005.14165, 1:3

    2020 arXiv

  11. [19]

    Christian WF Mayer, Sabrina Ludwig, and Steffen Brandt. 2023. Prompt text classifications with transformer models! an exemplary introduction to prompt-based learning with large language models. Journal of Research on Technology in Education, 55(1):125--141

    2023

  12. [20]

    Daniela Occhipinti, Serra Sinem Tekiroglu, and Marco Guerini. 2023. Prodigy: a profile-based dialogue generation dataset. arXiv preprint arXiv:2311.05195

    2023 arXiv

  13. [21]

    Wasu Top Piriyakulkij, Volodymyr Kuleshov, and Kevin Ellis. 2023. Active preference inference using language models and probabilistic reasoning. arXiv preprint arXiv:2312.12009

    2023 arXiv

  14. [22]

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. 2023. Direct preference optimization: Your language model is secretly a reward model. Advances in Neural Information Processing Systems, 36:53728--53741

    2023

  15. [23]

    Laria Reynolds and Kyle McDonell. 2021. Prompt programming for large language models: Beyond the few-shot paradigm. In Extended abstracts of the 2021 CHI conference on human factors in computing systems, pages 1--7

    2021

  16. [24]

    Wentao Shi, Mengqi Yuan, Junkang Wu, Qifan Wang, and Fuli Feng. 2024. Direct multi-turn preference optimization for language agents. arXiv preprint arXiv:2406.14868

    2024 arXiv

  17. [25]

    Yifan Song, Da Yin, Xiang Yue, Jie Huang, Sujian Li, and Bill Yuchen Lin. 2024. Trial and error: Exploration-based trajectory optimization for llm agents. arXiv preprint arXiv:2403.02502

    2024 arXiv

  18. [26]

    Alex Tamkin, Kunal Handa, Avash Shrestha, and Noah Goodman. 2023. Task ambiguity in humans and language models. In The Eleventh International Conference on Learning Representations (ICLR)

    2023

  19. [27]

    Jiaqi Wang, Enze Shi, Sigang Yu, Zihao Wu, Chong Ma, Haixing Dai, Qiushi Yang, Yanqing Kang, Jinru Wu, Huawen Hu, and 1 others. 2023. Prompt engineering for healthcare: Methodologies and applications. arXiv preprint arXiv:2304.14670

    2023 arXiv

  20. [28]

    Xingtong Yu, Chang Zhou, Yuan Fang, and Xinming Zhang. 2024. Multigprompt for multi-task pre-training and prompting on graphs. In Proceedings of the ACM Web Conference 2024, pages 515--526

    2024

  21. [29]

    Eric Zelikman, Georges Harik, Yijia Shao, Varuna Jayasiri, Nick Haber, and Noah D Goodman. 2024. Quiet-star: Language models can teach themselves to think before speaking. arXiv preprint arXiv:2403.09629

    2024 arXiv

  22. [30]

    Eric Zelikman, Yuhuai Wu, Jesse Mu, and Noah Goodman. 2022. Star: Bootstrapping reasoning with reasoning. Advances in Neural Information Processing Systems, 35:15476--15488

    2022

  23. [31]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  24. [32]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.